Tensile tests were performed to study the strength, modulus and elongation at break of the rayon Elium^® composite and to compare the respective results with the available natural fibre (NF) composite counterparts.

The incorporation of rayon instead of NF was seen to enhance the strength and toughness of the composites compared to that of natural fibre composites, as the breaking point elongated to a much higher value (Table 1) using the same manufacturing technique, as conducted by the authors [17,20]. It was shown that rayon fibre has higher elongation and lower modulus than flax fibre [21].

ANOVA analysis for the tensile strength of the above-mentioned composites is displayed in Table 2, which includes the variance of the tensile strength, i.e., between groups (BG) and within group (WG). SS refers to the sum of square, and df stands for degree of freedom. The F is the ratio of the BG mean square (MS) to the WG mean square. For all three composites, 5 samples were considered; therefore, there were 15 samples in total. If the p-value of the F-test is less than 0.05, as reported here, the difference between the average tensile strength from one type of composite to another at a 95% confidence level is statistically significant. Table 3 shows the ANOVA analysis of the tensile modulus; similarly, the P-value was calculated to be less than 0.05, indicating a statistically significant difference between the average tensile modulus from one composite to another at the confidence level of 95%.

It was detected that the rayon fibre treatments with acetylation, 3-methacryloxypropyltrimethoxy silane (MPS) and 3-aminopropyltriethoxy silane (APS) were found to reduce the tensile strength by 2.1–35% and modulus by 2–24% of the resulting composites; however, the elongation slightly increased only for MPS-treated rayon fibre unsaturated polyester composite, as measured by one of the authors before [22]. The transfer of stresses goes from matrix to fibre; thus, the fibre–matrix interface is very important but was not achieved in these investigations. As the tensile properties were more fibre-dependent, this explains the ineffectiveness of rayon fibre chemical modifications.

As illustrated in Figure 1, no indication of fabric delamination was detected in the cross-sectional image, and the mode of fracture was fibre breakage. The sites of fractures are shown in Figure 1a and fibres broke along the tension direction are displayed in Figure 1b. Good interfacial adhesion was observed between the rayon fibres and the Elium^® matrix.

The three-point bending strength and modulus of rayon fibre-reinforced Elium^® composite samples were measured in order to evaluate the properties and compare them with those of natural (ramie and jute) fibre Elium^® composite systems, as well as rayon fibre epoxy composites, fabricated via the same production method. The stress–strain curves of rayon fibre-reinforced Elium^® composite samples are displayed in Figure 2, and the average flexural strength and modulus are tabulated (Table 4). The values are 93.5 MPa and 5.6 GPa for strength and modulus, respectively, which are in the same range as rayon fibre epoxy composite (as described in the previous work) [23]. Moreover, 0&90 woven jute fabric Elium^® was found to show lower strength (7%) and modulus [20].

It was shown that the chemical treatment of rayon fibres did not lead to the enhancement of the flexural properties of their respective composites. For instance, γ-aminopropyltriethoxysilane (APTES) was used to modify the rayon fibre surface, and, in some composites, a silane-coupling agent was used for incorporation into the epoxy system; however, no improvement was recorded for the flexural performance of the developed rayon epoxy composites [23]. In another investigation by the author [22], 3-methacryloxypropyltrimethoxy silane (MPS) and acetic anhydride were employed for silane and acetylation treatments of the rayon surface to produce treated rayon unsaturated polyester composites, which did not provide positive results for bending modulus and strength.

It was observed that the bending samples did not fracture even after 7% bending strain; instead, very tiny cracks were observed on the tensile and compressive sides of the samples. On the tension and compression surfaces of the bent samples, the matrix became whiter and less transparent due to the breakage of polymer chains (Figure 3a,b). The cross-sectional views of the bent samples were detected to illustrate no signs of delamination or breakage, as depicted in Figure 3c.

Tabulated ANOVA analysis (Table 5 and Table 6) shows the variance of flexural properties for the rayon Elium^® and jute Elium^® composites. The variance of flexural strength calculated between the groups and within groups (each group contains the result of five samples) demonstrated that the p-value equals 0.05. This implies that the difference between the flexural strength of these two composites was relatively significant, at slightly less than a 95% confidence level. This value could be explained by the fact that the bending test is more matrix-dependent, and, as the same matrix was used to make these two composites, the results can partly overlap. However, a p-value of less than 0.05 was obtained from the F-test of the modulus values, which indicates a statistically significant difference between the average flexural modulus of these two composites at a confidence level of 95%.

Besides, in order to have a better understanding of the performance of the bio-based composite in an actual part, first finite element (FE) simulation of the bending scenario was calibrated and validated with the experiment, and subsequently, the materials model was applied to a car door panel for predicting the behavior of the component under flexural loading.

The whole flexural setup was designed according to the experiment, as shown in Figure 4, and a materials model was adopted as per the experiment test result. The width, length and span of the sample were 20 mm, 80 mm and 64 mm, respectively, and it was modelled as a homogeneous shell. The supports and the load, which are presented as a cylinder with a diameter of 5 mm, were modelled as rigid. The modulus was obtained from the experiment and density, and Poisson’s ratios were 1.29 g/cm³ and 0.37 [24,25], respectively. Coulomb friction was used to model the friction between surfaces with a coefficient equal to 0.2 [26]. Polynomial materials modelling was assumed to model the plastic constitutive behavior. The bottom cylinders were fixed in all directions and rotations, whereas the top cylinder was fixed in the x and z directions and could only move downwards (along the y direction). The mesh type of the sample was a 2D shell quad with an element size of 1 mm, and the number of elements was 1440. The parameters above were found to be optimal in terms of computational time and accuracy convergence after performing some preliminary investigations of the element size. The FE model with the mesh and the color representation of the von Mises stress subjected to the bending load is displayed in Figure 4a,b and a vertical displacement of 18 mm was registered in the center of the sample.

A polynomial function of degree two was used to estimate the plastic deformation of the sample and is obtained as where σ is the stress (in MPa); x is the instantaneous strain; a_k is the strain amplitude, and k is the strain exponent. Polynomial expression fitted the experimental results by using the strain amplitudes of a₂, a₁ and a₀, which were employed as approximately 8, 54 and 10, respectively, and the strain exponent of k = 2. All of these parameters were extracted with a coefficient of determination (R2) of 99%.

A comparison between FE model results in terms of force-displacement and the experimental measurements is depicted in Figure 5. Generally, a good agreement was obtained between the experimental result and the FE simulation. The small deviation in the plastic region is due to the fact that the average Poisson’s ratio used for this simulation was the value obtained from the literature, which could not demonstrate the precise amount. This validated the developed model and enabled its utilisation for the prediction of the bending performance of the real-life component.

The simulation enabled the assessment of the displacement-force performance and the deformation of the car door panel made of the same developed material in a bending scenario.

The swift car door panel [27] was used to simulate the bending of the part, which was modelled with the same materials model used for the bending test validation, under a flexural load in the shape of a sphere measuring 100 mm diameter. A friction coefficient of 0.2 was applied for the interaction between surfaces. The sphere placed in such a way to bend the middle of the door panel and its position from the bottom right corner of the door panel was 337 mm and 284 mm in the x and z directions, respectively. The dimension and mesh of the door interior are depicted in Figure 6. The FE model of the whole assembly, i.e., the sphere and the door panel is also shown in Figure 7. All the parameters were adjusted in accordance with the materials model developed in the flexural test experiment. The door panel was fixed from all the edges and the displacement to the Y direction was enabled in the assembly. When the sphere went down into the middle of the component, the maximum load was obtained at the deflection of around 100 mm (Figure 8b). However, the deflection could be much lower if the panel was supported by the door exterior. It is worth mentioning that the von Mises stress obtained (~86 MPa) was close to that of the bending test scenario (Figure 8a). Eventually, the prediction of the deflection-force behavior of the car door panel bent by the ball is shown in Figure 9. At around 102 mm deflection, a maximum load of approximately 20 kN could be recorded. This sort of materials modelling and structural simulation could contribute to the prediction of the performance and deformation of the automotive interior components, which can be fabricated using bio-based composites.

The DMA test was also performed to investigate the viscoelastic properties of rayon fibre Elium composite and to compare them with those of NF polymeric composites investigated by the authors, as shown in Figure 10.

The storage modulus (E′) was measured at 5.8 GPa at 30 °C, which is in good agreement with the flexural modulus as obtained from the three-point bending test at room temperature. The glassy state region lasted until 80 °C wherein glass transition phase was initiated, as illustrated in the E′ curve. The glass transition zone continued until about 100 °C, with a swift continuous drop in E′.

It can be discovered that the glass transition temperature of rayon fibre-reinforced Elium^® composite (107.7 °C), as obtained from the loss factor peak (tan (d)), is higher or comparable with that of rayon epoxy systems and other natural fibre Elium^® counterparts. The standard deviation of the T_g value for the composite is shown in parentheses in Table 7.

The whole mechanical results and viscoelastic analysis demonstrate the great potential of the developed bio-based composite for the interior applications, and structural simulation can be performed to initially predict the mode of deformation and expected force in the component.